Guildford BABCP conference: discussion on the Dodo assertion - all good depression treatments are equally effective (fifth post)
Last updated on 4th August 2011
I wrote yesterday about "Supershrinks and pseudoshrinks". The first panel discussion, that I then went to on the main conference's second day, overlapped a bit into this territory. It addressed the proposition "All bona fide psychological treatments for depression are equally effective" with a couple of experts supporting the notion and a couple opposing it. The abstract read "In Alice in Wonderland the Dodo famously proclaimed that "everybody has won and all must have prizes" This spawned the Dodo bird hypothesis pertaining to psychological treatments for emotional disorders which states that all bona fide interventions will be equally efficacious and enduring due to their combination of non-specific and specific ingredients. The Dodo bird hypothesis has proved provocative and in this debate we explore the arguments on both sides as they relate to clinical depression, taking in different schools of psychotherapy, meta-analytic data, different severities and subtypes of depression, and the emerging literatures on personalised treatments, case formulations and transdiagnostic approaches. Two speakers for, and two against, the motion that treatments are equally effective will present their views and will have the right to reply to what the others say. After each talk members of the audience will be invited to ask 2 or three questions and at the end it will be opened up for audience discussion and questioning of the speakers. The debate will end with an audience vote."
The Dodo bird hypothesis was apparently first proposed back in 1936 in a paper comparing different forms of psychodynamic therapy. Probably more people became aware of this issue through with Luborsky's famous 1975 paper "Comparative studies of psychotherapies. Is it true that "everywon has one and all must have prizes"?" (sic) with its abstract stating "Tallies were made of outcomes of all reasonably controlled comparisons of psychotherapies with each other and with other treatments. For comparisons of psychotherapy with each other, most studies found insignificant differences in proportions of patients who improved (though most patients benefited)." The overlap with yesterday's exploration of "Supershrinks and pseudoshrinks" emerges as the abstract goes on to comment "Our explanations for the usual tie score effect emphasize the common components among psychotherapies, especially the helping relationship with a therapist." In a later paper - "Do therapists vary much in their success? Findings from four outcome studies" - Luborsky revisited this territory, finding that "Success rates of psychotherapists were compared across each of four treatment outcome studies, with results indicating: considerable difference between therapists in their average success rates; considerable variability in outcome within the caseload of individual therapists; little support for the widely held view that certain therapists are best for certain kinds of patients; and variations in success rate typically have more to do with the therapist than with the type of treatment."
What did the experts speaking to this topic say? Much what you'd expect them to say! Steve Hollon, in his usual impassioned rapid-fire way, argued for the superiority of cognitive-behavioural therapy. His arguments however were almost entirely based on the very impressive work of Ehlers, Clark and their colleagues in their various developments of improved CBT treatments for PTSD, for social anxiety and for panic disorder. I've already visited this territory in the earlier post " ... Rolls Royce therapy and Anke Ehlers on PTSD". In a question, I commented to Steve that his arguments didn't apply to the topic of this debate which was specifically about the treatment of depression. I didn't also point out that his arguments might apply to CBT results using Ehlers, Clark et al's interventions, but - as shown by Lars-Goran Ost's systematic review - this doesn't necessarily extend to CBT for anxiety more generally. Alex Wood made an argument for the Dodo bird hypothesis. Part of his point was that it is very possible for somebody to cherry pick particular research studies to back up their argument that this or that particular therapy is better than others for depression treatment. However this cherry picking is precisely what good meta-analyses "iron out" with their overview of all relevant research - see, for example, the classic 2008 Cuijpers et al paper "Psychotherapy for depression in adults: a meta-analysis of comparative outcome studies" with its comments that "Although the subject has been debated and examined for more than 3 decades, it is still not clear whether all psychotherapies are equally efficacious. The authors conducted 7 meta-analyses (with a total of 53 studies) in which 7 major types of psychological treatment for mild to moderate adult depression (cognitive-behavior therapy, nondirective supportive treatment, behavioral activation treatment, psychodynamic treatment, problem-solving therapy, interpersonal psychotherapy, and social skills training) were directly compared with other psychological treatments. Each major type of treatment had been examined in at least 5 randomized comparative trials. There was no indication that 1 of the treatments was more or less efficacious, with the exception of interpersonal psychotherapy (which was somewhat more efficacious; d = 0.20) and nondirective supportive treatment (which was somewhat less efficacious than the other treatments; d = -0.13). The drop-out rate was significantly higher in cognitive-behavior therapy than in the other therapies, whereas it was significantly lower in problem-solving therapy. This study suggests that there are no large differences in efficacy between the major psychotherapies for mild to moderate depression."
It's worth pointing out that a subsequent Cuijpers et al paper - "Interpersonal psychotherapy for depression: a meta-analysis" - showed that "Ten studies comparing IPT and other psychological treatments showed a nonsignificant differential effect size of 0.04". It's very clear that IPT (along with CBT, BA, etc) is an effective treatment for depression, but once more the dreaded Dodo bird strikes ... IPT is not significantly better than other evidence-based psychotherapy approaches for depression. I'm a huge fan of the work of Pim Cuijpers and his colleagues - see, for example, the website "Psychotherapy: randomized controlled and comparative trials". We were fortunate to have one of his colleagues speaking at this debate - Professor Gerhard Andersson from Linkoping and the Karolinska in Sweden. He made a series of good points, not least the concerning comment that all evidence-based psychotherapies for depression might be equally effective, but one could also argue that they are equally ineffective - see, for example, "The effects of psychotherapy for adult depression are overestimated: a meta-analysis of study quality and effect size".
So what should we do with the very high probability that good psychotherapies for depression are currently all pretty much as effective (or ineffective) as each other? The fourth speaker, Ed Watkins, argued that there might be value in seeing depression as a heterogeneous condition and developing a module based treatment where interventions are chosen more for underlying processes than for the general depression diagnosis. That's OK, but I'm not holding my breath. For now, I would back the Cuijpers et al suggestion for a stepped care approach. See "Stepped care for depression in primary care: what should be offered and how?" and its comments "Stepped-care approaches may offer a solution to delivering accessible, effective and efficient services for individuals with depression. In stepped care, all patients commence with a low-intensity, low-cost treatment. Treatment results are monitored systematically, and patients move to a higher-intensity treatment only if necessary. We deliver a stepped-care model targeting patients with depression. The first step consists of "watchful waiting", as half of all patients with a depressive episode recover spontaneously within 3 months. The second step, guided self-help, is the key element of the stepped-care model. Guided self-help, especially when offered through the internet, is effective and cost-efficient. The third step consists of brief face-to-face psychotherapy. Finally, in the fourth step, longer-term face-to-face psychotherapy and antidepressant medication might be considered. Patients are monitored by one person, a care manager, who is responsible for the decision to step up to the next treatment and for continuity of care. The different treatments within the stepped-care model are evidence-based. Data on cost-effectiveness of the full model are still scarce, but we recently demonstrated that the incidence of new cases of depression and anxiety could be halved by introducing stepped care. Effects of web-based guided self-help could be enhanced by incorporating them in a stepped-care model." Good stuff. I also believe that results can vary considerably between therapists (rather than therapies) and that feedback systems should be put in place to track developing outcomes and make them available (along with average expected progress charts) to clients & therapists session by session - see yesterday's post on "Supershrinks and pseudoshrinks".
Now it was time for conference keynotes and further symposia - see tomorrow's post "Fathers & child anxiety, and more on couple therapy".